在Jessie中使用新的Linux內核

• 在 Docker 中使用 cfs_quota 控制CPU，致使內核崩潰：
  

CPU: 27 PID: 89704 Comm: exe Tainted: G         C    3.16.0-4-amd64 #1 Debian 3.16.7-ckt25-2+deb8u3
Hardware name: Dell Inc. PowerEdge M630/0PHY8D, BIOS 1.4.5 10/26/2015
task: ffff88084b5df2b0 ti: ffff880855214000 task.ti: ffff880855214000
RIP: 0010:[<ffffffff8109f954>]  [<ffffffff8109f954>] check_preempt_wakeup+0xd4/0x1d0
RSP: 0018:ffff880855217e60  EFLAGS: 00010097
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000008
RDX: 0000000000000000 RSI: ffff880fe065b4b0 RDI: ffff881055b9ae00
RBP: ffff880dfdfc01c0 R08: ffffffff81610960 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000000 R12: ffff8810007fa250
R13: ffff88107f1d2f00 R14: 0000000000000000 R15: 0000000000000000
FS:  00007f5ed7b7d740(0000) GS:ffff88107f1a0000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000078 CR3: 000000068f84a000 CR4: 00000000001407e0
Stack:
 0000000000012f00 ffff88107f1d2f00 ffff880fe065b4b0 ffff88107f1d2f00
 ffff880fe065bb34 0000000000000246 ffff880fddde0ac0 ffffffff81094565
 0000000000012f00 ffffffff810969ea 00007fffffffeffd ffff880fe065b4b0
Call Trace:
 [<ffffffff81094565>] ? check_preempt_curr+0x85/0xa0
 [<ffffffff810969ea>] ? wake_up_new_task+0xda/0x160
 [<ffffffff8106698c>] ? do_fork+0x13c/0x390
 [<ffffffff81514579>] ? stub_clone+0x69/0x90
 [<ffffffff8151420d>] ? system_call_fast_compare_end+0x10/0x15
RIP  [<ffffffff8109f954>] check_preempt_wakeup+0xd4/0x1d0
 RSP <ffff880855217e60>
CR2: 0000000000000078

• Docker信息以下（Bug說明 #13940）：

# docker info
Containers: 99
 Running: 14
 Paused: 0
 Stopped: 85
Images: 49
Server Version: 1.12.0
Storage Driver: aufs
 Root Dir: /docker/aufs
 Backing Filesystem: xfs
 Dirs: 402
 Dirperm1 Supported: true
Logging Driver: json-file
Cgroup Driver: cgroupfs
Plugins:
 Volume: local
 Network: bridge macvlan host null overlay
Swarm: pending
 NodeID: 526qb3beuge4zn9urapmu4uh6
 Is Manager: false
 Node Address: 192.168.22.33
Runtimes: runc
Default Runtime: runc
Security Options:
Kernel Version: 3.16.7-ckt25
Operating System: Debian GNU/Linux 8 (jessie)
OSType: linux
Architecture: x86_64
CPUs: 48
Total Memory: 62.82 GiB
Name: shd-docker-03
ID: MFNF:NBGM:2VYN:HLA4:XRWT:EBFI:LZTV:VHVX:5VA6:BAJC:JNXB:AHHG
Docker Root Dir: /docker
Debug Mode (client): false
Debug Mode (server): true
 File Descriptors: 112
 Goroutines: 211
 System Time: 2016-09-10T02:46:40.356550459+08:00
 EventsListeners: 2
Registry: https://index.docker.io/v1/
Insecure Registries:
 127.0.0.0/8

• 挑當前最新 LongTerm 版本：
  

• 下載、編譯、打包 內核過程記錄以下：

aptitude -y install fakeroot kernel-package libncurses5-dev build-essential
lftp -c 'pget -n 30 https://cdn.kernel.org/pub/linux/kernel/v4.x/linux-4.4.20.tar.xz'
tar -axf linux-4.4.20.tar.xz
cp /boot/config-$(uname -r) linux-4.4.20/.config
make menuconfig
# General setup  ---> Control Group support  ---> Group CPU scheduler  （全勾選） 
make -j 24 deb-pkg
ls ../*.deb

• 然而不幸的是，使用新內核沒法啓動（慘不能睹）：
  

• 原來升級內核也要講究順序的：

wget 192.169.23.233:2333/{linux-image-4.4.20_4.4.20-1_amd64.deb,linux-libc-dev_4.4.20-1_amd64.deb,linux-firmware-image-4.4.20_4.4.20-1_amd64.deb}
dpkg -i linux-libc-dev_4.4.20-1_amd64.deb
dpkg -i linux-firmware-image-4.4.20_4.4.20-1_amd64.deb
dpkg -i linux-image-4.4.20_4.4.20-1_amd64.deb
vi /etc/default/grub
update-grub
reboot
uname -a

• 4.4 啓動過程卡且慢，ssh登陸後手動操做感受到卡頓。回退到 4.1 版本則正常~